BMR Record 1970/89

c.4

•

•

Record No. 1970/89 053890

¥

Barometric Heighting - An Assessment

of the Accuracy Achieved during

Reconnaissance Gravity Surveys in Australia

by

F. Darby

I I I I I I I I I I I I I I I I I I I I

1970/89

BAROMETRIC HEIGHTING - AN ASSESSMENT

OF THKACCURACY"ACHIEVED-nURING'

RECONNAISSANCE GRAVITY SURVEYS IN AUSTRALIA

by

F. Darby

I I I I I I I I I I I I I I I I I I I I

CONTENTS

SUMMARY

INTRODUCTION

NETWORK ACCURACY

ACCURACY. BETWEEN BENCH.MABKS .. '.

INTERNAL.STANDARDDEVlATION

BAROMETER DRIFTS

DOUBLY READ ELEVATION INTERVALS

BAROMETER PAIR DIFFERENCES

CONCLUSIONS

RECOMMENDATION

~EFERENCES

FIGURES

1. Typical flight plan for 1 :250,000 area

2. Correlation of standard deviation of network adjustment with mean elevation interval

3. Differences between measured and known elevation dUferences on one loop

4. Comparison of known and measured elevation intervals

5. Correlation between height interval from ftxed point and measured elevation interval

6. Barometer drifts

7. Barometer drifts

8. Doubly read elevation intervals

9. Barometer differences

1

1

3.

3.

3

4

4

4

6

7

.1 I :1 I ·1 .1 I I I II I I I I I I I I I I

SUMMARY

The Bureau of Mineral Resources has currently completed a reconnaissance gravity survey of approximately 65% of Australia with a station density of approximately 1 per 50 square miles. The elevations for these stations have been obtained by barometric techniques. The results of these surveys indicate that erratic behaviour of the barometers and, at this stage,.unknown (but probably climatic or mathematical). systematic variations contribute to the errors in the results. With the techniques employed on the reconnaissance graV'ity surveys it appears unlikely that the. standard deviation of the adjustments to the survey network can be better than 3-4 feet. Trial surveys are recommended to investigate the possibility of using barometers reliably on more detailed surveys.

I I: I~

.1 I I I I I I I I I I I I I I I I

'~/ ". " ,

.. INTRODUCTION

The Bureau of Mineral Resources, Geology and Geophysics has currently completed a reconnaissance gravity survey of approximately 65% of Australia. Stations hay.e:been established at a density of approx­imately one per 50 square. miles. The elevations for these stations have been obtained by. barometric techniques. The survey techniques have undergone several modifications since the inception of this survey, but a fairly standard method has been in operation for the past few years. Figure 1 shows a typical flight plan for a standard 1 :250,000 map sheet and it can be seen that a strong'network if1J built up which is usually well controlled by accurate third order level bench marks.

From all these surveys a large amount of information has been accumulated as to the accuracy of barometric levelling techniques. Since 1965 the results of these surveys have been reduced by computer (Bellamy, Lodwick. & .. Townsend, .1970) and large. quantities of statistical data has been generated. This report presents some of the data, which have mainly been obtained from the 1968 sury~y (with which the author was most closely. connected), but data are also Included from the 1965, 1966 and 1967 surveys. The report is by no means comprehensive but it will indicate the expected accuracies that are likely to be obtained from such surveys.

The accuracy of the elevations on helicopter gravity surveys controls' the accuracy of the final Bouguer anomalies~ The statistics presented here indicate what this accuracy is likely to be. The formulae used in the height reductions have been discussed elsewhere (Bellamy & Lodwick, 1967) and are not repeated here.

NETWORK ACCURACY

As outlined previously each.loop (Fig....1).iorms. part of a network controlled by fixed heights. In the .. flying of. a loop the cell centre and tie point must not be located on opposite sides of a bench mark traverse. Under. these conditions areas (or segments) contained by bench mark traverses

. can be considered to be independent for computation purposes, and errors cannot be propagated into adjacent segments.

During computation each segment is computed three times through a least squares network adjustment phase:

(it)

(iii)

"'i'"{' . • , .... !1)~" nr .: , " :, ~'. I .. "..!.t) , : fT11

with only one fixed elevation in the network, (thfls give~ the flnternal st:aLndla:rd dev i1ai.U on) ;

wUhhalfof the fixed elevaUons 11n the netwolrk;

with all fixed elevations in the netwOlrk (this gl1v8&'l the external standalfd deviation),

In (U) the diffe'rences between the computed values and the tr'Ul~ valu.es at the bench marks which wel!"e not fixed are used to compute the f01recast standalfd deviation.

Table 1 gives the internal and external standard devi2l.Uons together with the maximuni adjustmerits. Also listed in Table 1 are the number of loops in each segmenty the maximum elevation difference on one loop between cell centlfe and tie pointy the average elevation d11.ffelfence of all loopsy and the numbelf of loops with elevation differences of ovelr 500 feet.

. ,', ..

In Figure 2 the average measur'ed:'elevation difference. is plotted against 'the internal and external standard deviations., for each segment, It is seen , that there is an increase in standard deviation with increase fln mean elevation difference (except for segmentDy which has a large percentage of loops with an elevaUon difference of over 500 feet). From this correlation it can be estimated that a standard deviation of less than 3 feet will not be very common even in relatively flat areas.

TABLE 1 1968 N.S,W~

Seg. Extel!"nal S.D, .Internal S .. D. Mean No. Max. No. of loops Elev. Loops diff. ,with elev,

SD MA SD MA diff. on ovel!" 500 ft Loop

A 11. 98 ' 35.02 5.85 ' 17.81 496 48 2868 17

B 20.80 56.71 10.33 32.93 965 88 4169' 46.

C 10.54 40.06 4,77 11.68 375 53 2021 13 ;

D 18.72 44.06 14.34 39.37 633 39 1890 18

F 9,04 23.66 6.24 14.60 338 102 1943 24

G 7,86 17.92 4.42 11.78 53 62 225 0

H 6,45 19.77 4.21 11.10 57 97 285 0

I 5.64 14.91 5.11 12.65 61 59 270 0

J 4048 12.33 2.46 7.13 74 36 220 0

K 5.37 12.68 2,53 5.28 38 35 149 0

L 5.50 18.63 4,08 18.71 76 85 289 0

I I I I I I II ,_I

I I I I I I I I I I I 1/

I I I I I I I I I I I I I I I I I I I I

....... '.'

ACCURACY BETWEEN BENCH MARKS

DUIC'ing the 196J~ sUIrvey in New South W:ffiles~ Queensland~ and Western AlIlstlraUa many loops contained .more than one bench mark. Comparison of the measured elevation differences with the true elevation differences on these loops gives an estimate of the accuracy that ca."} be expected in measured elevation intervals. These differences (sign.ignored) are plotted as a histogram on Figure 3 .. Of these measured intervals, 72Yo are accurate to 7 feet or better and 86% are accurate to 10 feet or better.

The majority of the dtlferences are plotted against the true elevation interval on Figure 4. No really Significant conclusion emerges from this as the majority of the elevation intervals. measured were comparatively small (less than 500 feet). However, there is a tendency for larger errors to be associated with larger elevation intervals.

INTERNAL STANDARD DEVIATION

As stated plr'eviollsly the 'internal standard deviation is obtained from the least sq1,utares adjustment of the network using only one fixed elevation. In most parts of Austlr'alia the elevation range in an individlllal segment is smaU~ but in the coastal segments of the 1968 survey in New South Wales and Queensland a wide range in elevation was encountered (Table 1). Figure 5 shows the height of the bench marks above or below the fllxed node and the difference between the computed and true elevations for these stations. It can be seen that there is a defilnite correlation between height intel'Val and error.

BAROMETRIC DRIFTS

During the flying of a loop a base barometer is read at the commencement point of that loop. The difference in reading between the field and base barometers are compared at the beginning and end of each loop and the discrepancy between these two diffelr'ences is ·called the baJrometric drift. Histograms of these drifts for the 1965, 1966~ 1967, and 1968 surveys are presented in Figure 6 and the total drifts are presented in Figure 7 together with the normal distribution curve having the same mean and standard deviation. It can be seen that the standard deviation of these drifts is 5.64 feet. Any drifts in excess of 12 feet are rare and can generally be considered suspect mis:readings or an indication of barometer malfunction.

DOUBLY READ E LEV AT ION INTERVALS

During the course of the more recent surveys stations have been established along the eastand west margins of all 1:2501000 map sheets (Fig. 1). These stations aIC'e occ':lpied twice on different days. The primary objective of these stations is to assist the Department of National Mapping topographic .. map selfies. However 1 the double occupation means that certain elevation intervals are measured twice. This gives an independent check on the repeatability of barometric height interval measurements. Figu:rre 8 is a histogram of the differences between these doubly read elevation intervals.

BAROMETER PAIR DIFFERENCES

Dulfing a detailed g:rravUy survey at Gosses Bluff in 1969 a pair of balfometers were read simultaneously at each survey point. The differences in barometer readings at each point are shown in Figure 9 and it can be seen that there is a wide scatter in the differences. The largest would appear .. to be dUlringthe middle of the day (i.e. the time of maximum temperature), but the sample is not large enough and the results not consistent enough to conflllfm this. The differences commonly range over 0.3.0 millibars on anyone day. This means that independent elevation differences measured by either of these barometers could differ by over 10 feet. This erratic behaviour of the barometers could account to a large extent for the barometric drift discussed earlier.

CONCLUSIONS

The main conclusions derived from the preceeding discussion are:

1. The erratic behaviour of barometer pairs causes both barometric 'drift' ovelr a loop and also errors in elevation differences between stations.

2. There appears to be a linear lfelationship between the average elevation difference between tie points and the standard deviation of the adjustments to the netwol1"k.

3. A systematic erlr'Olf of up to 100 feet occurs over areas with total '\. elevation dtffell"ences of 3000 feet.

• • • • • • • • • • I I II I

-I I II I I

The following alf€ the more important stCBLUStiCS reveCBLled~

1. The minimum startdalfd deviation of the adjustments to a barometlfic network is 3=4 feet with an as~ociated maximum 1iJl.djustment of 9=12 feet.

2. Comparison with knowri elevation differences indic1ill.tes that there is a mean error of about 6 feet.

3. The standard deviation of the barometric drifts is approximately 5.5 feet.

4. Elevation differences which are read twice have a mean difference of about 6 feet.

It is felt that in areas of low relief during stable weather conditions most of the observed errors in the network can be accounted for by the differential behaviour of the barometers with time~ The response of the barometers both to abuse in the field and to temperature variations may be important in this regard.

In areas of high relief the physical characteristics of the barometers should remain the same (if the barometers are in good condition) and the factors which become important are:

(a) different weather patterns and associated pressure conditions between adjacent valleys and ranges;

(b) different temperature conditions between the base barometer and field barometer.

In rugged terrain it is probable that better results would be obtained if the elevation difference on loops were restricted to between 200 and 300 feet.

Over the maj ority of Australia mountainous conditions do not occur and so we expect and usually obtain elevations accurate to within 10 feet. This is accurate enough for a reconnaissance survey. In mountainous environments 1 however, the rapidly changing climatic conditions only permit an accuracy of about 40 feet.

It is important to emphasise that generally it is only the changing climatic conditions that cause the major inaccuracies (greater than 10 feet) in barometric levelling. This is exhibited in coastal vicinities where on-and off-shore breezes often cause systematic errors to occur in the barometric elevations.

-6-

RECOMMENDATION

It was concluded that the elevations obtained by barometric levelling are generally accurate enough for reconnaissance gravity surveys, but for more detailed surveys, when Bouguer anomaly features of less than 2 mgals could be important, the present system 'Would be inadequate.

It. is therefore recommended that experimental barometer surveys be conducted over known elevation ranges in

and

(a) flat inland areas

(b) mountainous areas

(c) coastal area.

Combinations of multibase and.multifield barometer techniques should be tried in an endeavour, to. improve the accuracy of the elevations. Recording of field temperatures will also be important. If a technique can be evolved to obtain accuracies of 2-3 feet it may be a more economical and faster way of conducting semi-detailed gravity surveys.

I I I I I I I I I I I I I I I I I I I I.

I I I I I I I I I I I I' 'II II II II II I;

I I

-7-

REFERENCES d'

BELLAMY, C.J., and LODWICK, G.D., 1967 - Reduction of barometric networks and field gravity surveys. Pub!. Computer Centre, Monash Univ.

BELLAMY, C.J., LODWICK, G.D., and TOWNSEND, D.G., 1970 - BMR gravity reductions, storage and retrieval system. Bur. Miner. Resour. Aust. Rec. (in prep.). '971-7 iRe-c)

•

- • I

~ 1 I 1 i~

I> ._ I •

IBaaed On 06,-1&0)

TYPICAL FLIGHT PLAN FOR 1'250.000 AREA

LEGEND

• Bench mark

• Gravity station

CC Cell centre

TP Tie point

@ Dcrubly read stati~"~ an sheet margin

- - - - - - - - .- - - - _____ "65-__

--------------------

CO)

en UI I

N .,. N J:o

Z o i= cS: :;

15

~ 10 o 0:: .... < .... OILI ZILI <11.1. .... -UI 5 ..J

~ 0:: ILl

o

0-

o o·

o

o 0

.... ~

(0) O~----------~---------------------------------

20 Z 0

fi :; ILl 0

15 0 0::.-< .... Ow 2:11.1 < .... .... --(/)

10 ..J <t Z 0 0:: ILl .... )( ILl

5 0

0 0 100 200 . 300

MEAN

400

ELEVATION

0'--

CORRELATfON OF. STANDARD DEVIATION OF NETWORK

ADJUSTMENT WITH MEAN ELEVATION INTERVAL

1968 HELICOPTER GRAVITY SURVEY. NSW ANO QLD

(b)

500 600 700 800 900 1000 INTERVAL ON LOOP (FEET)

. /

:!! (i) c: ::0 I'll

N

--------------------

CO) CJ) (JI ,. N ~ o >

"

> u Z

. ILl ~ 0 ILl II: IL.

30.

20

10

DIFFERENCES BETWEEN MEASURED AND KNOWN ELEVATION

DIFE~ENCES O~ONE LOOP_

5 ". 10

1968 HELICOPTER GRAVITY SURVEY N SW. ~LD. AND WA

178 OBSERVATIONS

15

FEET

MEAN 6-04 FEET

25 30. 35

--------------------

' ... ' ....

1600

~ fi >. ~ 1&1

.~ 600 o z ~

.: 400

200

',:'

." '~.

COMPARISON OF KNOWN AND MEASURED ELEVATION INTERVALS .

. ::'.

,-. . ", -,'

+ 1-

1968 HELICOPTER GRAVITY SURVEY NSW, CLD, AND WA

+

. +.

+

+

+

+ .

+ +

+ +

f' + + + + ;

+

.... + + .. ' • +

+ .•

+ +

+.+ + '-f + +

+ 1-

+

+

+

+.' . + +. .

+

1-

:.-----

+

+

:~. -(i) c:

* ~ 0~·-------L----~~L-------~----~~~~~~~~~~0~~~~5~-L~~+I~0------~--------L---~-+-2SL-------+~30-------+~35· ~ -25 -20 . -15 -10 -5 1-

MEASURED- KNOWN ELEVATION INTERVAL (FEET) ~

- - - - -

~ ..,

·+100

+80

+60

+40

~ +20

z o t= ~ IIJ ..J .., IIJ ;:) a: ~

I z

.0

fi > .., ..J IIJ

o .., a: ;:) fI) 4( .., ~

o

-20

-40

-80

-100

o o

o

- - - - - - - - -

o

CORRELATION BETWEEN HEIGHT INTERVAL FROM FIXED POINT AND' MEASURED ELEVATION INTERVAL

1968 HELICOPTER GRAVITY SURVEY SEGMENT A,8,C,D, AND F NSW AND OLD

(CALCULATED FROM INTERNAL STANDARD DEVIATION COMPUTATION)

..

o o .. ...

0·0

o 0 o

-3000

•

o· •

•

o· o

o

-2000

eo o

o

o

o

o

o 0

o • . 0 0 0 0 00 .. a.: or o

o o

0 0 0 0 0 o 0 o 0

o o ·0

00

o·

o

o

o

o o

o

ELEVATION DIFFERENCE FROM FIXED NODE

o

o ,

o

o

- - - - - -o

o

o

o o

o

o '0

o

. ." -G) C

.::0 m 01

II - -80-

60- ,... u C .. ::J r7

40- :! ~

20 -'

01

-20

120-

190-

80-

,.. u c·

60 - .. ::J r7 to .. ~

40-

20 -

o - .

- - - - - -1~65 H~LlCOPTERGRAVITY SURVEY

729 oblervations Standard deviation: 5.79 feet Meon:+O.97 feet

1

-16 1

. -12 1

-8 1 I I.

-4 0 +4 Barometer drift

1 +8

1967 HELICOPTER GRAVITY SURVEY

1049 observations Standord deviation: 5.50 feet Meon: ... 7'feet

I

-16 I

-12 I

-8 I I I

. -4 0 '.4

Barometer drift

, I .8

- - -~AROMETER DRIFTS

1

+12

I +12

1 1 1 +16 +20Feet -16

I +16 Feet

I -16

-

1

-12

I

-12

- - - - - - -1966 HELICOPTER GRAVITY SURVEY

I -8

I I I -4 0 +4 .

Barometer drift

595 obser.vations Stondard deviation: 5.71 feet

. Mean:.0.48feet

I

·8 1

·.12 1

+16 Feet

196L_.ti~J..J.~QPTER 'GRAVITY SURVEY

I I 1 -8 -4 0 +4

Boram!ter drift

1076 obsenotions . Stondard deviation: 5.35 feet Mean :-039 'feet

I

,+8 , .12

~ -Ci) I C

+16 'Feet ::u :1"1

(J)

r - -

G') en <.n I

rJ CD J>

- -

,.. u c .. ;, CF .. ...

280 -

240-

200-

160-

~ 120-

80 -

40-

- -

0- -I

-20 I

-16

-

1 -12

- - - -- - - - -- -BAROMETER DRIFTS

1965,1966, 1967 ond 1968 HELICOPTER GRAVITY SURVEYS

1 -8

I 1 1 -4 0 +4

Barometer drift

1 +8

3449 OBSERVATIONS

STANDARD DEVIATION: 5.64 FEET

MEAN: +0.69 FEET

/ Normal c:urve

I .20 Feet

-

." -G') C ::0 IT1

.......

-

--------------------

. 25

. I ~ ..

.::. '20 ,'.":

'.J.,

t; Z &&.I ::» o III

15

e: 10

5

'. DOUBLY READ ELEVATION INTERVALS

19*38 HELICOPTER GRAVITY SURVEY. NSW, QLD, AND WA

167 OBSERVATIONS

MEAN 6-14 FEET

.. -' - ~ -. -~ '" .

" .~-.

O~--~----~~------.---__ --~--~~~--~~~-L--~~--~~~ 5 ~ ffi ~ ~ ~. o

DIFFERENCE IN DOUBLY READ ELEVATION INTERVALS (FEET)

'"" -Ci) c: ::0 ",

Q)

I· 1 1 1 1 1 1 1 I 1 I I I I I I I I I 1

+0'10

o 4ft IE -0'10 C III -0,20

~ -0030 ~

• -0·40

-0'110

all-e-..

-0.60 -I----.-~__,~.....,......;......___'T"'-~..;..,_-.,._....., 8 9 10 II la IS 14' 15 18 11 18

" 0 U R S

+0010 ae-e-ea

~-~: \;~r. ~ -0'20 ~

• -o·so

FIGURE 9 If-e-ee

, Iii f , iii I .. ea' 10 II la IS 14 III 18 IT

-0040 -I-......,IO~..,II.--..." .... a--.IS .'-1 "';"lr~.--r/i/''''''''l2r, -.-T'/'s-. 'T1"-'T/~-"T/'e-"T:1-"T/~---'/~ .

:to-20

+0'/0 CIt II:

0 c III - -0·10 ~ ~

• -0,20

-0,30

-0040 9

+0·20

." ..0'10 II: ·c 0 III

,~ -0'10 ~

- -0,20 .. ' -0' SO

. -0040

• +O·ao

'" II: +0010 c III 0 ~

~ -0,/0

• -0,10

-0,30 I

" 0 U R S

1-1-ea a,:'-e.

I J:Y V

I I I I I , , 10 II 12 IS 14 II Ie IT II la la IS 14 III 18 11 /I ,

" 0 U R S

4-1-ea 'a-1-"

-\ -V\ . ~

I I I I I I I

/0 II 12 IS 14 III Ie IT II la I • ·10 " 12 IS 14

" o U R 8

1-1-e.

~.j a 10 II la IS 14 III 18 IT It

(" 0

TO accompany R«ord No. 197rJ/89

a-1,-ea.

~-\jv

, , I I I 19 14 III IS IT II

U R S

BAROMETER DIFFERENCES

GOSSES BLUFF, NT 1969

BAROMETERS 7&1 AND 1173

I II

F &3/B2-&OA